Photocatalytic Activity of the Oxide Layer Formedon NiTi Surface through Thermal Oxidation Process

Kouta Sakamoto+1, Kento Yokoi+1, Aki Saito+2 and Naofumi Ohtsu+3

Instrumental Analysis Center, Kitami Institute of Technology, Kitami 090-8507, Japan

In the present study, we attempted to prepare a photocatalytic titanium dioxide (TiO2) layer on nickel-titanium (NiTi) alloy surface througha simple thermal oxidation process in air. At 723K, an amorphous TiO2 layer including a slight amount of Ni was formed on the surface. Above873K, the TiO2 layer crystallized into a rutile phase. At 1023K, a complex oxide of NiTiO3 was also formed. The methylene blue degradationtest for evaluating photocatalytic activity showed that the formed TiO2 layers act as photocatalyst under ultraviolet (UV) light illumination, andits activity is superior to that of the surface layer formed by oxidizing a pure Ti substrate. The adhesion strength of the surface layers formed at723K on NiTi alloy was higher than that of a commercially available TiO2-coating. We conclude that thermal oxidation as a surface modificationtechnique is expected to make NiTi alloys give photocatalytic activity. [doi:10.2320/matertrans.M2014114]

(Received March 31, 2014; Accepted May 14, 2014; Published July 25, 2014)

Keywords: nickel titanium, thermal oxidation, photocatalytic activity

1. Introduction

A nickeltitanium (NiTi) alloy has been used as a materialfor dental and medical devices such as root canal files andcatheters, due to its shape memory property and super-elasticity.1) In medical practice, these devices must generallybe sterilized in order to remove microorganisms attached ontheir surfaces, and high-pressure steam or dry-heat treatmentshave been used as sterilization processes. However, the riskof infectious disease transmission from the devices persists,even after the aforementioned processes. It is considered thatthis risk can be reduced if the NiTi alloy itself possessesan antimicrobial activity. Thus far, researchers have madeefforts to increase the antimicrobial activity of metallicsubstrates.26)

Our research group has focused on photocatalytic TiO2 asa possible antimicrobial medium. Photocatalytic TiO2 hasexcellent antimicrobial activity under ultraviolet light (UV)illumination, and is not modified in the process.710)

Furthermore, TiO2 is a chemically stable, nontoxic material.To coat TiO2 on metallic substrates, chemical and physicaldeposition techniques such as plasma spraying and sputteringhave been used.11,12) However, the adhesion strength of thecoating using these techniques is comparatively low, ingeneral. Thus, these processes are not considered suitable formedical devices. Compared to those deposition techniques,the adhesion strength of surface layers formed by oxidizing ametallic surface is high due to its gradual interfacial structure.Anodic oxidation process is possible candidate to prepareTiO2 layers. Wong et al. reported that the oxide layer couldbe formed on NiTi surfaces through alternating currentanodization, resulting in a loose surface layer with largecracks.13) Such loose surface layer may injure a human body,causing inflammation. Therefore, this technique is notsuitable for medical devices.

A simple thermal oxidation process in an atmosphereincluding oxygen is suggested to be an alternative technique

for forming TiO2 layers on these alloys. It is well known thatthe oxidation reaction’s elemental selectivity in the thermaloxidation of an alloy is dominated by the standard freeenergy for the formation of metallic oxides. For instance,the values at 298K for TiO2 and NiO are 889.5 and211.7 kJmol¹1, respectively. This selectivity implies that theTi contained within the alloy is preferentially oxidized whenheated it in air, resulting in the formation of a TiO2 layer.Vojtěch et al. reported that a TiO2 layer was formed on NiTialloy by heating it in air at 803K,14) and Wu et al. reportedthat a TiO2 layer was prepared on NiTi alloy by heating it inair at 723K.15) These reports revealed that TiO2 layers couldbe formed through heating. However, in those studies, thepurpose of the formation of TiO2 layers was to create brier forNi ion diffusion to the human body, and the photocatalyticactivity of the layers was hardly discussed. In this study,therefore, oxide layers were formed on NiTi surfaces througha thermal oxidation process, and their characteristics wereinvestigated in detail. Furthermore, the photocatalyticactivities of surface layer were estimated by using amethylene blue (MB) degradation test.

2. Material and Methods

NiTi disks (55.07 at% Ni, NEC TOKIN, Japan), 15mm indiameter and 2mm in thickness, was used as a substrate. Thesubstrate was mechanically polished using emery papers upto #1200, and the disk was then ultrasonically rinsed inethanol for 600 s. The thermal oxidization was conductedusing a conventional electronic furnace (FO-100, YamatoScientific, Japan) in an ambient atmosphere. The substratewas placed in the furnace, heated at oxidizing temperaturesranging from 573 to 1023K for 18 ks. Then, the furnace wasturned off and allowed to cool to room temperature, with thesample inside.

The surface morphologies of the oxidized substrates wereobserved using scanning electron microscopy (SEM; JCM-5000 Neo Scope, JEOL, Japan), with an acceleration voltageof 10 kV, using secondary electron image. The crystallinitywas investigated using X-ray diffractometry (XRD; New D8ADVANCE, Bruker AXS, Germany) in the Bragg-Brentano

+1Graduate Student, Kitami Institute of Technology+2Undergraduate Student, Kitami Institute of Technology+3Corresponding author, E-mail: nohtsu@mail.kitami-it.ac.jp

Materials Transactions, Vol. 55, No. 8 (2014) pp. 1332 to 1336©2014 The Japan Institute of Metals and Materials

http://dx.doi.org/10.2320/matertrans.M2014114

geometry using Cu K¡ radiation. The chemical state anddepth profile of the surface layers were characterizedusing X-ray photoelectron spectroscopy (XPS; PHI5000VersaProbe, ULVAC-PHI, Japan) with monochromatizedAl K¡ radiation (h¯ = 1486.6 eV). For the XPS measure-ments, the diameter of the X-ray probe was approximately100 nm, and the photoelectron take-off angle was set at 45°.The binding energy was corrected using the C 1s peakcorresponding to surface adsorbed hydrocarbon (284.8 eV).Elemental depth profiling was done using an Ar ion gun withan acceleration voltage of 4 kV. The etching rate estimatedfrom the SiO2 layer was approximately 0.2 nm s¹1. Theadhesion strength of the surface layers on the substrate wasestimated from the critical load to exfoliate the oxide layerusing nanolayer scratch tester (CSR-2000, Rhesca, Japan).The surface layers were scratched using a diamond styluswith 200 gmm¹1 spring constant and tip radii of 5 µm. Thelateral rate, amplitude and loading rate were 10 µm s¹1,100 µm and 2.50mN s¹1, respectively.

The photocatalytic activity of the oxidized substrate wasestimated using a methylene blue (MB) degradation test. Priorto the test, the specimens were immersed in a polypropylenevessel containing 5 dm3 of 10 µmol cm¹3 MB aqueoussolution, after which the vessels were placed in the dark for86.4 ks, in order to complete the adsorption of MB molecules.The aqueous solution was thereafter refreshed, and irradiatedwith light of a 370 nm wavelength and an intensity of1mWcm¹2 for 10.8 ks, using an LED. The MB concentrationwas evaluated every 1.2 ks during the illumination bymeasuring the absorbance of MB at 664 nm, using a UVvis spectrometer (UV-2400 PC, Shimadzu, Japan). Theevaluated concentration was plotted versus the illuminationperiods, and MB degradation rates (µmol cm¹3 s¹1) werecalculated from the slope of the resulting line.

3. Results and Discussion

3.1 Characteristics of the oxidized NiTi surfaceMorphological images of thermally oxidized NiTi alloy

surface with various temperatures, observed by SEM, areshown in Fig. 1. An SEM image of an untreated NiTi alloysurface is also shown for comparison. The surface morphol-ogy of the alloys oxidized at 573K (Fig. 1(b)) and 723K(Fig. 1(c)) are similar to that of the untreated alloy surface(Fig. 1(a)). The formation of the surface layer could not beconfirmed from the SEM images. On the other hands, atransition in surface morphology, towards higher granularity(Fig. 1(d)), was observed at oxidation temperatures above873K, and the exfoliation of specific areas was observed onthe surface after oxidizing at 1023K (Fig. 1(e)). The grainsobserved on the surface after oxidizing at 1023K were muchlarger than those observed after oxidizing at 873K.

The XPS spectra of the Ti 2p, O 1s and Ni 2p regions,obtained from NiTi substrates oxidized at 723 and 873K, areshown in Fig. 2. In both temperatures, the binding energy ofthe Ni 2p3/2 peak is 856.2 eV and small satellite peaksassociated with Ni 2p peaks were observed. These character-istics in Ni 2p region agree with those of NiO.16) Similarly,in the case of O 1s, the spectral shape for 723K almostcoincides with that for 873K. The binding energy of the O 1s

peak is 530.2 eV, which agrees with that for metallic oxidestate.16) On the other hand, slight chemical shift in Ti 2pspectra was observed between the 723 and 873K surfaces.The binding energy of the Ti 2p3/2 peak for 873K is458.5 eV, which almost agrees with that of TiO2,16) whereasthe energy for 723K is slightly lower than that for 873K.The atomic ratio of Ni to Ti ([Ni]/[Ti]) for 723 and 873K,calculated from the spectral intensity of Ni 2p and Ti 2p,were 1.1 and 0.4, respectively. Wu et al. analyzed the bindingenergy of nanostructured TiNiO compounds with various[Ni]/[Ti] ratios using XPS, revealed that the energy ofTi 2p3/2 was shifted toward low value with the increase of[Ni]/[Ti] ratio.17) Considering from their report, we con-cluded that the slight shift of Ti 2p peaks was originated fromthe difference in the [Ni]/[Ti] ratio in the topmost surface.Combined, these results indicate that the surface chemicalstate of the oxidized NiTi alloys is composed of TiO2 withsmall amounts of NiO.

XRD patterns of the oxidized NiTi alloys at varioustemperatures were measured using a Bragg-Brentano geom-etry, and are shown in Fig. 3. In the case of oxidationtemperatures below 723K, oxides peaks were hardlyobserved. After oxidizing at 873K, peaks corresponding tothe rutile phase of TiO2 appeared in the patterns. Foroxidizing temperatures between 873 and 1023K, theintensities of the TiO2 peaks drastically increased. Thischange indicates the extensive development of the oxidelayer. Furthermore, small peaks corresponding to the

(a)

20µm

(c)(b)

(d)

20µm

5µm

20µm 20µm

(e)

5µm

20µm

Fig. 1 SEM images of NiTi alloy substrate without treatment (a), and withthermal oxidization at temperatures of 573K (b), 723K (c), 873K (d), and1023K (e).

Photocatalytic Activity of the Oxide Layer Formed on NiTi Surface through Thermal Oxidation Process 1333

complex oxide NiTiO3 were also found in the pattern afteroxidation at 1023K. The XRD results demonstrated that acrystallized TiO2 layer can be formed on NiTi alloys throughthermal oxidation by heating above 873K, although the useof too high of a temperature such as 1023K, results in theformation of a complex oxide. The NiTiO3 is formed byreacting NiO and TiO2, and its standard free energy decreaseswith increasing temperature.18) When employing compara-tively low temperature, NiO formed in surface layer. The NiOreact with TiO2 when raising the temperature, therebyforming NiTiO3.

XRD measurements cannot detect amorphous compounds.Thus, in order to confirm the existence of amorphous TiO2 onthe surface, we measured a depth profile of the oxidized NiTialloy surface. Elemental depth profiles for NiTi alloysoxidized at 573 and 723K are shown in Fig. 4. In the caseof 573K, oxygen was detected only in the topmost surfaceregion, and decreased abruptly away from the surface. Thisresult indicates that a distinct TiO2 layer was not formed onthe surface. On the other hand, after heating at 723K, oxygendiffused into the alloy up to about 100 nm from the surface,and the atomic ratio of O to Ti was approximately 2 to 1 inthis region. Concomitantly, a Ni content of approximately10 at% was found in the shallower region. Firstov et al.reported that the inward diffusion of oxygen and the outwarddiffusion of Ti from the interior occur simultaneously whenoxidizing NiTi alloy at the temperature above 673K.19) Thesimilar diffusion of Ni also occurs during the oxidization;however, its speed is slower than that of Ti, thereby formingTi-oxide leyer.19) We considered that the Ni remaining in thesurface layer diffused toward both outward and inwarddirections after forming Ti-oxide because Ni atoms diffuseeasily through defective TiOx.19,20) Consequently, the Ni-richregions at the topmost surface and beneath the oxide layerformed on NiTi alloy.

3.2 Photocatalytic activity of the oxide layerThe formation of the crystallized or amorphous TiO2 layer

was confirmed on a NiTi alloy oxidized at temperaturesabove 723K. This result lets us expect that the oxidized alloyshould act as a photocatalyst. The photocatalytic activityunder UV light illumination as estimated by the MBdegradation test is shown in Fig. 5. Test results obtainedusing two types of control samples consisting of first, pure Tioxidized at various temperatures, and second, a TiO2 layerformed using a commercially available TiO2-coating regent(ST-K211, Ishihara Sangyo, Japan), are also shown in thefigure. In the case of the NiTi alloy, the degradation rate ofthe MB was almost zero at an oxidation temperature of

844849854859864869874879884889894Int

ensi

ty (

Nor

mal

ized

)

Binding energy, Eb/eV

453458463468Inte

nsit

y (N

orm

aliz

ed)

Binding energy, Eb/eV

522527532537542Inte

nsit

y (N

orm

aliz

ed)

Binding energy, Eb/eV

Ti 2p O 1s

Ni 2p

Ti 2p3/2

Ti 2p1/2

Ni 2p1/2Ni 2p3/2

SatelliteSatellite

723 K

873 K

723 K

873 K

723 K

873 K

Fig. 2 XPS spectra of Ti 2p, O 1s and Ni 2p regions obtained from the NiTi surface oxidized at 723 and 873K.

20° 25° 30° 35° 40°

Inte

nsit

y (A

rb. u

nit)

Diffraction angle, 2θ

1023 K

573 K

723 K

873 K

Untreated alloy

TiO2 (rutile)NiTiO3

Fig. 3 XRD patterns of the oxidized NiTi alloy substrate measured byusing Bragg-Brentano geometry.

K. Sakamoto, K. Yokoi, A. Saito and N. Ohtsu1334

573K. The rate drastically increased for oxidation temper-atures greater than 723K, and the highest value was observedfor an oxidation temperature of 873K. The rate decreases ifthe sample is oxidized above 1023K. These results show thatNiTi alloys oxidized above 723K act as a photocatalyst underUV light illumination, although the activities are lower thanthat of a coating formed using a commercially availableregent. It should be noted that the value at 873K is slightlyhigher than that of 723K. This difference is probably due tothe crystallinity of the oxide layer. Furthermore, the usage oftoo high temperature deteriorates this activity because thepeeling off of the layer occurs.21)

Photocatalytic activity was also observed on the oxidizedpure Ti substrate (Fig. 5). However, for oxidation temper-atures above 723K, the rates are lower than those for

oxidized NiTi alloy surfaces. We conjectured that thissurprising result originates from the NiO present on thetopmost surface. Sreethawong et al. reported that the loadingof NiO enhances the photocatalytic activity of TiO2 becauseNi2+ oxidation state in NiO can trap photo-generatedelectrons, thereby suppressing the recombination of thephoto-generated electrons-hole pairs.22) The trapped electronis released immediately; therefore, the activity is notdeteriorated. Consequently, we consider that the Ni2+ presentin the TiO2 layer enhanced the photocatalytic activity of theoxidized NiTi surface. On the other hand, it is well-knownthat Ni ion released from a metallic surface causes allergicreaction and cytotoxicity. This effect makes us hesitate to useoxidized NiTi as medical devices, although NiO in thetopmost surface is beneficial in the view from photocatalyticactivity. However, the concentration of Ni existing in thetopmost surface was below 20 at%, which was lower thanthat of Ni contained within NiTi alloy. Furthermore, weconsider that a NiTi alloy with photocatalytic activity wouldbe used in medical devices that are repeatedly used clinically,such as root canal files and catheters. In these applications,the amount of Ni ion released from the surface would betrivial because these devices are not implanted for a long timein a human body. On the basis of these considerations, weconcluded that the harmful effect from the NiO present intopmost surface is negligible.

3.3 The adhesion strength of the oxide layer on the NiTiA merit of thermal oxidization process is the ability to

form surface layers having high adhesion strength. Toconfirm this, we estimated the adhesion strength of the oxidelayer on the NiTi alloys using a nanolayer scratch tester. Thecritical loads of the oxide layer for oxidation temperature of723 and 873K are compared with that of commercial-reagentTiO2-coating in Fig. 6. We couldn’t measure the strength ofthe layer oxidized at 1023K because the layer had alreadyexfoliated after the oxidation (Fig. 1(e)). The critical load of

0

20

40

60

80

100

0 100 200 300

Ato

mic

con

cent

rati

on (

%)

Depth, d/nm

0

20

40

60

80

100

0 100 200 300A

tom

ic c

once

ntra

tion

(%

)Depth, d/nm

573 K 723 KNiTiO

NiTiO

Fig. 4 The elemental depth profiles of the NiTi surface oxidized at 573 and 723K.

0

0.05

0.1

0.15

0.2

0.25

0.3

Rea

ctio

n ra

te c

onst

ant

of M

B d

egra

dati

on, c

/nm

ol c

m-3

s-1

Heating temperature, T/K

Untreated substrate

573 723 873 1023

Commercially available TiO2-coating

NiTiTi

Fig. 5 MB degradation rate of the NiTi alloy substrates oxidized withvarious temperatures. This result obtained from commercial-regent TiO2coating is also inserted. All error bars are mean « S. D.; n = 3.

Photocatalytic Activity of the Oxide Layer Formed on NiTi Surface through Thermal Oxidation Process 1335

the layer oxidized at 723K is higher than that oxidized at873K, which is more than double of the commercial-regentTiO2-coating. From the XPS profile (Fig. 4), the thickness ofoxide layer formed at 723K is about 100 nm, and the layerthickness probably increases with the increase of thetemperature. Kim et al. reported that the oxide layer of1.2 µm in thickness was formed on NiTi alloy when heating itat 873K.23) We therefore considered that the difference of theadhesion strength depending on the temperature is due to thethickness of the oxide layer.24) The surface layer of thesample oxidized at 723K showed a notable advantage in itsadhesion strength although its photocatalytic activity wasinferior to that of commercial-reagent TiO2-coating (Fig. 6).In medical practice, the peeling off of the surface layer isdirectly linked to the risk of the failure in medical operation;accordingly, the adhesion strength is the most significantproperty. Consequently, we believe that thermal oxidationusing a temperature of 723K is the optimum treatmentprocedure for medical safety to give NiTi alloys photo-catalytic properties.

4. Conclusions

We prepared photocatalytic TiO2 layers on NiTi alloysurfaces through a simple thermal oxidation process. TheTiO2 layers were formed by heating NiTi alloys in air attemperatures above 723K, and the crystallinity of oxidevaried from amorphous to rutile with the oxidation temper-ature. The prepared layers included small amounts of Ni inthe topmost surface regions, in a NiO chemical state. Theoxidized surface acted as a photocatalyst under UV lightillumination, and its activity was higher than that of oxidizedpure Ti substrates. The adhesion strength of the surface layerfor a 723K oxidation temperature was notably higher thanthat of TiO2 layers produced using a commercially available

reagent. In conclusion, thermal oxidation using a temperatureof 723K is an excellent technique to produce photocatalyticsurface layers having high adhesion strengths on NiTi alloysurfaces, and this technique is promising for use in medicalpractice.

Acknowledgement

The authors gratefully acknowledge Mr. M. Yamane fromKitami Institute of Technology for his support with XRDanalysis, and also acknowledge Mr. W. Saito, a graduatestudent in our laboratory, for his help in XPS measurements.This work was supported by a Grant-in-aid for ScientificResearch (C) (No. 24560841) from the Ministry of Educa-tion, Culture, Sports, Science and Technology (MEXT) ofJapan.

REFERENCES

1) O. Fukushima, T. Yoneyama, H. Doi and T. Hanawa: Dent. Mater. J. 25(2006) 151160.

2) V. Sharma, R. Yngard and Y. Lin: Adv. Colloid Interface Sci. 145(2009) 8396.

3) T. Tanabe, N. Okitsu, A. Tachibana and K. Yamaguchi: Biomaterials 23(2002) 817825.

4) H. Sashiwa and S. Aiba: Prog. Polym. Sci. 29 (2004) 887908.5) G. A. Sotiriou and S. E. Prasinis: Environ. Sci. Technol. 44 (2010)

56495654.6) R. A. Ahmed, S. A. Fadl-allah, N. El-Bagoury and S. M. F. G. El-Rab:

Appl. Surf. Sci. 292 (2014) 390399.7) A. K. Benabbou, Z. Derriche, C. Felix, P. Lejeune and C. Guillard:

Appl. Catal. B 76 (2007) 257263.8) D. M. A. Alrousan, P. S. M. Dunlop, T. A. McMurray and J. A. Byrne:

Water Res. 43 (2009) 4754.9) Y. Oka, W. Kim, T. Yoshida, H. Mouri, H. Urade, Y. Itoh and T. Kubo:

Int. J. Biomed. Mater. Res. Part B: Appl. Biomater. 86B (2008) 530540.

10) P. Evans, T. English, D. Hammond, M. E. Pemble and D. W. Sheel:Appl. Catal. A 321 (2007) 140146.

11) F. X. Ye, A. Ohmori and C.-J. Li: Surf. Coat. Technol. 184 (2004) 233238.

12) S. Takeda, S. Suzuki, H. Odaka and H. Hosono: Thin Solid Films 392(2001) 338344.

13) M. H. Wong, F. T. Cheng and H. C. Man: Appl. Surf. Sci. 253 (2007)75277534.

14) D. Vojtěch, J. Fojt, L. Joska and P. Novak: Surf. Coat. Tech. 204 (2010)38953901.

15) S. Wu, X. Liu, Y. L. Chan, J. P. Y. Ho, C. Y. Chung, P. K. Chu, C. L.Chu, K. W. K. Yeung, W. W. Lu, K. M. C. Cheung and K. D. K. Luk:J. Biomed. Mater. Res. 81A (2007) 948955.

16) Y. W. Gu, B. Y. Tay, C. S. Lim and M. S. Yong: Appl. Surf. Sci. 252(2005) 20382049.

17) Y. Wu, Y. He, T. Chen, W. Weng and H. Wan: Mater. Lett. 59 (2005)31063108.

18) K. T. Jacob, V. S. Saji and S. N. S. Reddy: J. Chem. Thermodyn. 39(2007) 230235.

19) G. S. Firstov, R. G. Vitchev, H. Kumar, B. Blanpain and J. V.Humbeeck: Biomaterials 23 (2002) 48634871.

20) N. Shevchenko, M.-T. Pham and M. F. Maitz: Appl. Surf. Sci. 235(2004) 126131.

21) C. H. Xu, X. Q. Ma, S. Q. Shi and C. H. Woo: Mater. Sci. Eng. A 371(2004) 4550.

22) T. Sreethawong, Y. Suzuki and S. Yoshikawa: Int. J. Hydrogen Energy30 (2005) 10531062.

23) K. S. Kim, K. K. Jee, W. C. Kim, W. Y. Jang and S. H. Han: Mater. Sci.Eng. A 481482 (2008) 658661.

24) N. Ohtsu, T. Takahara, M. Hirano and H. Arai: Surf. Coat. Tech. 239(2014) 185190.

0

10

20

30

40

50

60

70

80C

riti

cal l

oad,

F/m

N

Oxidized NiTi at 723 K

Oxidized NiTi at 873K

Commerciallyavailable

TiO2-coating

Fig. 6 The critical load of the oxide layer formed by oxidizing NiTi alloy.This result obtained from commercial-regent TiO2-coating is also insertedfor comparison. All error bars are mean « S. D.; n = 3.

K. Sakamoto, K. Yokoi, A. Saito and N. Ohtsu1336

http://dx.doi.org/10.4012/dmj.25.151http://dx.doi.org/10.4012/dmj.25.151http://dx.doi.org/10.1016/j.cis.2008.09.002http://dx.doi.org/10.1016/j.cis.2008.09.002http://dx.doi.org/10.1016/S0142-9612(01)00187-9http://dx.doi.org/10.1016/S0142-9612(01)00187-9http://dx.doi.org/10.1016/j.progpolymsci.2004.04.001http://dx.doi.org/10.1021/es101072shttp://dx.doi.org/10.1021/es101072shttp://dx.doi.org/10.1016/j.apsusc.2013.11.150http://dx.doi.org/10.1016/j.apcatb.2007.05.026http://dx.doi.org/10.1016/j.watres.2008.10.015http://dx.doi.org/10.1002/jbm.b.31053http://dx.doi.org/10.1002/jbm.b.31053http://dx.doi.org/10.1016/j.apcata.2007.01.039http://dx.doi.org/10.1016/j.surfcoat.2003.11.012http://dx.doi.org/10.1016/j.surfcoat.2003.11.012http://dx.doi.org/10.1016/S0040-6090(01)01054-9http://dx.doi.org/10.1016/S0040-6090(01)01054-9http://dx.doi.org/10.1016/j.apsusc.2007.03.055http://dx.doi.org/10.1016/j.apsusc.2007.03.055http://dx.doi.org/10.1016/j.surfcoat.2010.05.010http://dx.doi.org/10.1016/j.surfcoat.2010.05.010http://dx.doi.org/10.1002/jbm.a.31115http://dx.doi.org/10.1016/j.apsusc.2005.03.207http://dx.doi.org/10.1016/j.apsusc.2005.03.207http://dx.doi.org/10.1016/j.matlet.2005.05.044http://dx.doi.org/10.1016/j.matlet.2005.05.044http://dx.doi.org/10.1016/j.jct.2006.07.016http://dx.doi.org/10.1016/j.jct.2006.07.016http://dx.doi.org/10.1016/S0142-9612(02)00244-2http://dx.doi.org/10.1016/j.apsusc.2004.05.273http://dx.doi.org/10.1016/j.apsusc.2004.05.273http://dx.doi.org/10.1016/S0921-5093(03)00287-9http://dx.doi.org/10.1016/S0921-5093(03)00287-9http://dx.doi.org/10.1016/j.ijhydene.2004.09.007http://dx.doi.org/10.1016/j.ijhydene.2004.09.007http://dx.doi.org/10.1016/j.msea.2006.12.221http://dx.doi.org/10.1016/j.msea.2006.12.221http://dx.doi.org/10.1016/j.surfcoat.2013.11.038http://dx.doi.org/10.1016/j.surfcoat.2013.11.038